Cognitive function and the performance of a secondary, dual task may affect certain aspects of gait, but the relationships between cognitive function and gait are not well understood. To better understand the motor control of gait and the relationship between cognitive function and gait, we studied cognitive function and the effects of different types of dual tasking on the gait of patients with Parkinson's disease (PD) and controls, contrasting measures of gait automaticity and rhythmicity with other features. Patients with idiopathic PD (n=30; mean age 71.8 year) with moderate disease severity (Hoehn and Yahr Stage 2--3) were compared to age and gender-matched healthy controls (n=28). Memory and executive function were also assessed. In both groups, gait speed decreased in response to dual tasking, in a parallel fashion. For the PD group only, gait variability increased compared to usual walking. Executive function was significantly worse in the PD group, while memory was not different in the two groups. Executive function measures were significantly correlated with gait variability during dual tasking, but not during usual walking. These findings demonstrate that regulation of gait variability and rhythmicity is apparently an automatic process that does not demand attention in healthy adults. In patients with PD, however, this ability becomes attention-demanding and worsens when subjects perform secondary tasks. Moreover, the associations between executive function and gait variability suggest that a decline in executive function in PD may exacerbate the effects of dual tasking on gait, potentially increasing fall risk.

The objectives of the present study were to test the hypothesis that the dual-tasking effect on gait variability is larger in healthy older adults than it is in healthy young adults; that this effect is larger in idiopathic elderly fallers than it is in healthy older adults; and that the dual-tasking effects on gait variability are correlated with executive function (EF). Young adults and older adults who were classified as fallers and nonfallers were studied. Gait speed, swing time, and swing time variability, a marker of fall risk, were measured during usual walking and during three different dual-tasking conditions. EF and memory were evaluated. When performing dual tasks, all three groups significantly decreased their gait speed. Dual tasking did not affect swing time variability in the young adults and in the nonfallers. Conversely, dual tasking markedly increased swing time variability in the fallers. While memory was similar in fallers and nonfallers, EF was different. The faller-specific response to dual tasking was significantly correlated with tests of EF. These findings demonstrate that dual tasking does not affect the gait variability of elderly nonfallers or young adults. In contrast, dual tasking destabilizes the gait of idiopathic elderly fallers, an effect that appears to be mediated in part by a decline in EF.

Walking is generally viewed as an automated, over-learned, rhythmic motor task and may even be considered the lower-limb analog of rhythmic finger tapping, another automated motor task. Thus, one might hypothesize that walking would be associated with a simple rhythmic task like tapping rather than with a complex motor task like catching. Surprisingly, however, we find that among older adults, routine walking has more in common with complex motor tasks, like catching a moving object, than it does with tapping. Tapping performance, including both the average tapping interval and the variability of tapping interval, was not significantly associated with any gait parameter (gait speed, average stride time and stride time variability). In contrast, catch game performance was significantly associated with measures of walking, suggesting that walking is more like catching than it is like tapping. For example, participants with a higher gait speed tended to have lower times to first move when catching, better catching accuracy, and less catching errors. Stride time variability was significantly associated with each of the measures of catching. Participants with a lower stride time variability (a more steady gait) had better catching accuracy, lower time to first move, fewer direction changes when moving the cursor to catch the falling object, and less catching errors. To understand this association, we compared walking performance to performance on the Stroop test, a classic measure of executive function, and tests of memory. Walking was associated with higher-level cognitive resources, specifically, executive function, but not with memory or cognitive function in general. For example, a lower (better) stride time variability was significantly associated with higher (better) scores on the Stroop test, but not with tests of memory. Similarly, when participants were stratified based on their performance on the Stroop test and tests of memory, stride time variability was dependent on the former, but not the latter. These findings underscore the interconnectedness of gait and cognitive function, indicate that even routine walking is a complex cognitive task that is associated with higher-level cognitive function, and suggest an alternative approach to the treatment of gait and fall risk in the elderly.